Pier39 AI

Four months ago, I wrote about the day SwimOutlet.com handed us 100% of their traffic on Day One for our post-purchase product. Now I've got their CMO, Adam Hayim saying it himself, on camera. His words: "If you built something that your customers trust, you can't afford for any part of the journey, let alone the post-purchase moment, to feel like it's veering off of the brand." Full conversation here: youtu.be/39EEafjQmHQ?si=RD1F…

One month of Nash in the wild. Time to put some numbers behind it. 📈📊 In the last month, Nash: → Lifted carts by 7% → Increased AOV by 24% → 52% of shoppers who interacted with Nash closed their cart → 95% of carts Nash closed did so without a discount The 24% higher AOV is especially interesting. We're not just seeing Nash recover more carts. We're seeing shoppers who interact with Nash buy more. That means Nash isn't just rescuing a shopper who was about to leave. It's helping shoppers make a decision, find the right product, or see more value in the cart they're building. And Nash is doing that while preserving margin in 95% of the carts we close. One month in. The thesis is holding up. 🚀🚀 pier39.ai/nash

We ran an experiment with Nash. It started asking customers leaving the site 1/ what stopped them buying today, and 2/ what price they would have bought at. ~25% answered!! We thought it will be close to 5% "$100 and I'd take it right now." "Free shipping and I'm in." Nash can now close it on the spot inside a floor you pre-approve, or counter. We also stream every response live into the merchant's Slack. Their team watches and acts on it in real time. They love it!! Best week of building we've had. pier39.ai/nash

🕰️ Week 2 of posting one real conversation Nash had with a shopper. A shopper at a healthcare brand added something to their cart… then just sat there. It was a high-ticket items. Nobody buys these without thinking first. Nash stepped in and asked what was holding them back. “I am not sure about the right product for me.” That’s the whole objection. They’d already picked something out and still weren’t confident it was the right one. Every default in ecommerce says throw a discount at that moment. 15% off the wrong product is still the wrong product. So Nash offered a free 15-minute consultation with one of the brand’s experts. They took it. The brand got a real conversation with someone who was about to close the tab. I keep running into this. The shopper isn’t standing there doing math on the price. They’re looking at four products that look identical to them… with nobody around to ask.

Starting something I’ve wanted to do for a while. Every week I’ll post one real conversation Nash had with a shopper. First one: Shopper at a natural foods brand hits the checkout page… and stalls. Nash: “What’s holding you back?” Them: “Price is too high.” Easy move would have been to slap on 15% off, take the order, move on. But then you’re just discounting someone who was already almost buying. Nash offered the brand’s membership instead: $99/year. 20% off everything. Free shipping. Express fulfilment. They signed up. Discount is almost never the best lever.

🚨 Day 28 of Nash being Live --> A Report Card Every brand runs the same lost shopper playbook: 10% off → 15% off → 20% + free shipping …and hope they forgot why they left. Nash just watched 17,000 real shoppers hesitate in real time. The discount... pier39.ai/nash

Founder lessons I wish someone tattooed on my arm in early 2026: 1/ First 7-figure contract didn’t come from fancy demos.

You spent $50 to land a new customer on your website. They're hesitating at checkout. You have two options. A) Talk to them right there, understand the hesitation, and close the deal if it's within your margins. B) Send a blanket 20% discount by email three days later, if they even open it. A used to be impossible to scale. Not anymore.. Try Nash (nash.pier39.ai) Our latest launch video, MOSTLY true 🙂

Agents can now handle real chopping through a clean set of tools. We released three functions that cover the full checkout flow: find_stores, list_products, and create_nash_order. They work without API keys, redirects, or custom work per merchant. Here is how the stack operates.👇

@elonmusk , @finkd , @sama have all tweeted about nash.pier39.ai in the last 24 hours. See what the fuss is about 👇

1846: One department store couldn’t train every clerk to talk with every shopper. So they had the bright idea called the "fixed price tag". It scaled retail… and became the default for 180 years. Until today. We just killed it!

By far the best investment I’ve made as a founder was getting a sales mentor. I was winging enterprise sales last year. He’d sold a similar-ish product to the exact same ICP for 20+ years. He completely rebuilt my approach, turned me into a closer, and cut my learning curve by months (maybe years). Sales is now my biggest strength. Lesson: Get specific mentors for your real bottlenecks -- sales, fundraising, team, storytelling, whatever it is. The acceleration is insane.

Been thinking about this a lot lately: Everyone's hyped that agents will shop for us. Sure! But the part i keep coming back to is what happens to price once the agent shows up, not whether it shows up. The fixed price tag is 180 years old. It exists because no store could afford to have a different conversation with every single customer walking in. It was never the fair price, just the only one that scaled. We got so used to it we forgot it was a workaround. That constraint is gone now. A conversation that used to cost you a human's time costs pennies of inference. so the economics changed, not just the interface. Some stuff i actually believe at this point: 1/ The fight isn't for the agent's attention, it's for the first offer. once discovery collapses into one llm answer there's no second click to win back 2/ Blanket discounts are the real race to the bottom, not conversational commerce. Giving everyone 20% off gives money to the people who would've paid full price anyway 3/ Cart abandonment isn't a ui problem like everyone says. it's usually a pricing problem showing up as a ui problem 4/ The protocol matters way more than the chatbot. anyone can slap a chat widget on checkout. the actual hard part is a standard any agent can talk to without you integrating with them first. 5/ Agent identity should be something the buyer chooses to share, not something we assume for them. Wrote the long version of all this if anyone wants it: blog.pier39.ai/posts/agentic…

🌶️ take: 50%+ OF ALL UTILITY SHOPPING WILL BE DONE BY AGENTS by end of 2027 Q. What is utility shopping? A. Where the shopping outcome and not the experience is the point e.g. You need nappies at certain cadence. You need a flight to Chicago. You need the same protein powder you buy every month. i.e., The purchase is a chore and you want it done, well, and fast. A few use cases: 👇